CONCEPTS FOR Anthropogenic Ecotope Mapping (AEM)
Web Version 1.0 (alpha) February 2006
This document provides a conceptual basis and overview for
Anthropogenic Ecotope Mapping & Classification (AEM;
et. al. 2006). It is not intended as a stand alone document and is
intended solely for use as part of standardized training in the
AEM system provided by a qualified instructor.
Proper citation of this document:
Ellis, Erle C. 2006. Concepts for Anthropogenic Ecotope Mapping
(AEM) (Version 1.0). Ecotope.org. February 2006. The Ecotope Mapping
Working Group. <Date of Access> <http:/www.ecotope.org/aem/training/concepts/default.aspx>.
Table of Contents
Long-term ecological changes within
densely populated anthropogenic landscapes account for a growing share of
global environmental change. Measuring the causes and consequences of these
changes is challenged by their fine spatial scale and complexity (Ellis
et. al. 2006) Anthropogenic ecotope mapping (AEM)
was designed to support the measurement of these changes at the fine spatial scales
at which they occur, and to integrate these landscape measurements with data on
land management practices obtained directly from local land managers and ecological
measurements obtained directly by sampled measurements in the field.
AEM is a standardized ecological mapping
procedure designed explicitly for the high-resolution mapping of ecologically-distinct
features within densely populated landscapes from a combination of high spatial
spatial resolution (≤ 1 m) imagery and intensive fieldwork used to verify the mapping
and classification of all features.
Two scale-explicit standards are used in AEM, a
Level 1 procedure used for relatively rapid current and historical mapping
and an even finer-scale
Level 2 procedure used to map sample areas within Level 1 maps, allowing
the correction of maps for the finer-scale features left out of Level 1 maps.
Ecotope mapping is a multi-stage process:
- A regional analysis is used to identify
Level 1 (L1) sample cells for ecotope mapping.
L1 mapping is the central method for tying together the local, site-specific
ecological characteristics of landscapes with the regional characteristics.
1 ecotope features are mapped within the sample AOI's based on interpretation
of ≤ 1 m resolution imagery by a combination of direct image interpretation and
groundtruthing by experts with local knowledge.
- A highly detailed, fine scale,
Level 2 (L2) mapping is conducted in sampled small quadrats and/or features
Level 1 mapped areas to observe features and ecological details not measurable
Level 1 system, such as fine linear features, small structures, riparian
zones, field borders, isolated trees, and small patches of managed vegetation. Level
2 data is combined with
Level 1 data to estimate areas and amounts of fine-scale patterns.
- Regional analysis is completed using
Level 1 data supplemented by
Level 2 details.
Level 1: Sample Cell Mapping (L1)
Applied to sample cells selected by a Regional Analysis. Lower resolution direct
and ground-truthed interpretation of ≤ 1 m resolution imagery. Classification of
both IKONOS imagery & WW2 aerial photos should be possible.
L1 mapping is considered the standard mapping system, with
Level 2 mapping used to measure fine-scale features within
L1 mapped areas. Mapping is usually accomplished within standardized
sample cells (usually 500 m � 500 m square cells), with a single trained interpreter
capable of mapping > 1 km2 in < 30 days, including all fieldwork and data
processing. After mapping, sample cells may be reassembled to form maps of
different shapes and sizes.
Level 1 Feature Mapping Scales (minimum mapping dimension)
≥ 2 m with area ≥ 25 m2 for
≥ 5 m with area ≥ 25 m2 for
hard areal features.
≥ 10 m with area ≥ 100 m2 for
soft areal features.
Level 2: Map Sample Mapping (L2)
Applied to sampled quadrats and/or features within
Level 1 mapped sample cells. High-resolution direct and ground-truthed interpretation
of features in IKONOS imagery within quadrats sampled within
L1 mapped areas or other sample sites. May not be possible for historical
aerial photo interpretation at all sites.
Level 2 Feature Mapping Scales (minimum mapping dimension)
≥ 0.1 m for
≥ 2 m for
hard areal features.
≥ 5 m for
soft areal features.
Overview. The sequence and scale of feature mapping follows the relative precision
by which different types of land use and vegetation features can be identified in
1 m resolution imagery and by groundtruthing using boundaries recognizable to both
land managers and to ecologists in the field in the face of land use/land cover
- Map linear features.
- Map hard areal (polygon) features.
- After mapping linear and hard features, soft features remain in the interstices.
- Some small soft features are mapped automatically by being
"cut out" of the landscape
by enclosure within linear and hard features; these include some features that would
otherwise be too small to map as separate soft features.
- Map soft areal features with relatively clear edges such as management boundaries,
vegetation planting patterns, and recently burned areas.
- Map large soft vegetation features without managed or other clear edges within
any remaining soft features with area ≥1 ha, whenever contiguous vegetation cover
patches ≥ 30 m in dimension (area ≥900 m2) differ significantly from surrounding
cover. This scale rule limits vegetation cover mapping by image interpretation
and groundtruthing to larger, more consistently identifiable features and facilitates
the use of higher spectral resolution systems (e.g. Landsat) for vegetation classification
where vegetation cover varies more gradually.
Linear features have clear edges and length ≥4 times their width, such as roads, paths, ditches, canals, streams, and hedgerows.
Hard areal feature: hard areal (polygon)
features have clear edges and relatively homogenous interiors; examples are constructed,
barren, and water surfaces and rice paddies.
Soft areal feature: features with fuzzy
edges and variable interiors, such as crop plots and patches of trees.
Ellis E. C., H. Wang, H. Xiao, K. Peng, X. P. Liu, S. C.
Li, H. Ouyang, X. Cheng, and L. Z. Yang. 2006. Measuring long-term ecological
changes in densely populated landscapes using current and historical high resolution
imagery. Remote Sensing of Environment 100(4):457-473. [download]